pfm wrote:
Thanks all for replies. Please google for 'Fundamentals of Floating
Measurements and Isolated Input Oscilloscopes'. This is a Tektronix
application note. The situation to which I refer is is at the bottom
of page 4. I want to use an oscilloscope to check the PWM switching
chip on the primary side of an SMPS.

BobH

You do NOT want to do this. Leave the scope grounded and put an
isolation transformer on the device under test.
But you still DON'T want to hook the ground of the probe onto the
common node of the DUT. there's just too much common mode stuff
floating around.
You can get some information by using two probes and a dual-channel
scope with one channel inverted and summed. Hook the probe grounds
together, but to NOTHING else.
You have to turn down the scope gain until EACH channel can handle
all the common mode signals on the inputs. Often, that results in
a sensitivity too low to get much info from the difference trace.
And you'll see junk because the ground path is ill-defined.

The TEK A6902A isolated probe works well in this application.
You can hook the ground lead of the probe almost anywhere you want.

In a perfect world, you can get by doing dangerous things. But it
takes only one slip-up to smoke your probe, your scope, your DUT
and YOURSELF!!!